An llms.txt file is a proposed plain-text file that some site owners use to point AI assistants towards key pages or summaries, a bit like an extra signpost for AI readers. Most businesses don’t need one as a priority. It isn’t a proven ranking or citation lever, it isn’t an official web standard, and there’s no strong first-party evidence from Google, OpenAI, Microsoft or Anthropic that adding it improves visibility in AI answers. If your basics are weak, you’ll gain more by fixing your AI welcome sign, making sure your site is in Google’s phone book and Bing’s phone book, publishing clear answer content, and tracking citations properly.
llms.txt is an optional signpost, not a core requirement
An llms.txt file is a text file placed at the root of your site, usually at /llms.txt. It’s meant to help AI readers find important pages, policies, product information, or summaries quickly.
That sounds useful. The problem is evidence. There’s still no public first-party proof from the major AI platforms that this file meaningfully improves how often your business is cited or recommended.
For most established companies, this means one thing. Treat llms.txt as optional housekeeping. Don’t treat it as a growth lever.
HelloWorld’s position is simple. If a practice is cheap and harmless, you can keep it in the hygiene bucket. If it isn’t backed by evidence, don’t let it displace work that clearly matters, such as accessible pages, strong answer content, comparison pages, author pages, reviews, roundups, business listings, and monthly crawler-access monitoring.
Your business usually does not need llms.txt before the basics are sorted
If your goal is to be cited by AI assistants, your first gains usually come from stronger foundations. That means being easy to access, easy to understand, and easy to trust.
Your AI welcome sign matters more
robots.txt, your AI welcome sign, is the small file that tells AI readers whether they’re invited in. If that sign blocks an answer crawler, or if your content is blocked at the server or CDN level, an llms.txt file won’t rescue you.
This has become more important since Cloudflare changed defaults for some new zones. Cloudflare said on 15 September 2026 that its Training and Agent crawler categories are blocked by default on ad-bearing pages for new zones, while Search stays allowed. It also classed ChatGPT-User as an Agent, and said a crawler with several roles gets the most restrictive rule. The practical effect is clear. A site can have a perfect AI welcome sign and still return a 403 block to ChatGPT-User at the CDN layer, according to Cloudflare’s product update on 15 September 2026.
That’s why access checks now carry the same weight as file checks.
Being in the right phone books matters more
Google’s phone book and Bing’s phone book still matter, but in different ways. Google Search Console launched its Generative AI performance report worldwide on 31 August 2026. Google said the report shows impressions inside AI Overviews and AI Mode by page, country, date and device. It does not show clicks or queries, and it does not split the two features, according to Google Search Console documentation dated 31 August 2026.
For Bing and Copilot, Bing Webmaster Tools remains important. Microsoft put its AI Performance report into public preview on 1 February 2026. Microsoft said it includes total citations, average cited pages, grounding queries, and page-level citations. Query-to-page mapping was added on 23 March 2026, according to Bing Webmaster Tools documentation.
For ChatGPT search, Bing is no longer enough on its own. OpenAI’s search changed on 4 September 2026 so that ChatGPT search runs on OpenAI’s own index, called Labrador, with search-result scraping as a fallback and Bing used mainly in Deep Research. The practical lesson is simple. Presence in Bing’s phone book no longer guarantees presence in ChatGPT search.
An llms.txt file does not solve any of that.
Clear answer content matters more
AI assistants cite pages that answer a question directly in visible page copy. That’s why answer-first writing matters. It’s also why HelloWorld doesn’t sell extra markup as a magic shortcut.
Google’s own guidance has also moved away from the idea that you need to rewrite pages into robotic chunks for AI. Google’s generative AI optimisation guidance states that rewriting copy “for AI” is unnecessary, and that content chunking is not required because its systems understand multi-topic pages. That is Google’s public guidance, and it supports a plain approach: write clear pages for humans first.
If you want more on this, see how to measure your AI search citation performance and what schema markup is, and whether it matters for AI search.
There is no strong evidence that llms.txt improves AI visibility
The safest way to explain llms.txt to a business owner is this. It may help some AI readers understand your preferred pages. It has not been shown, by first-party platform evidence, to be a reliable way to win citations.
That matters because many technical ideas sound useful in theory. Your budget should still follow what can be checked.
Today, the measurable signals available to site owners come from platform tools such as Google Search Console’s Generative AI report, Bing Webmaster Tools AI Performance, and traffic tools such as GA4. On 13 May 2026, Google Analytics 4 added a default AI Assistant channel for referrals from ChatGPT, Gemini, DeepSeek, Copilot and Grok. Google said this channel excludes Google AI Overviews and AI Mode, is not retroactive, and does not include Perplexity and Claude on the published list. That means the number is useful, but it is a floor, not a full KPI.
None of those first-party tools reports llms.txt as a performance lever. None provides a before-and-after metric for it. That’s why we keep it in the hygiene category, if used at all.
llms.txt can still be worth adding in a narrow set of cases
There are cases where adding the file is reasonable. The key is to keep expectations low.
Large sites with messy navigation can use it as a tidy signpost
If your site has thousands of pages, several subfolders, or duplicated paths, a short llms.txt file can act as a neat pointer to your most useful public pages. This is most relevant for software companies, merchants with large catalogues, and service firms with sprawling resource centres.
Highly regulated businesses can use it to point towards current policies
Online pharmacy, finance, legal and healthcare-adjacent businesses often need AI readers to find the right policy, credentials, and contact pages quickly. In those cases, a simple file that points to current public pages may reduce ambiguity.
It still doesn’t replace visible trust signals on-page. Real bylines, honest dates, sourced figures, pricing pages, delivery pages, returns pages, and review signals still do the heavy lifting.
Content teams can use it as internal discipline
Sometimes the main value is operational. Creating an llms.txt file forces your team to decide which pages are truly important. That can expose weak pages, outdated pages, or missing commercial pages such as pricing and alternatives.
If you’re missing those pages, fix that first. For many businesses, a strong alternatives page will move the needle more than a new text file.
What to put in an llms.txt file if you choose to use one
If you decide to add the file, keep it simple. Think of it as a short reading list for AI readers, not a place for tricks.
Include only public, useful pages
Point to pages that help a buyer, researcher or assistant verify who you are and what you sell. Good examples include your homepage, service pages, pricing page, alternatives page, product category pages, support centre, returns policy, about page, author pages, and contact page.
Use plain text and clear labels
Do not overcomplicate the file. Use short labels and full URLs. Keep it readable by a person as well as a machine.
Do not use it to hide weak foundations
If a page is blocked, thin, out of date, or not trusted, listing it in llms.txt won’t make it authoritative. Your visible content still matters most.
A practical example of a sensible llms.txt approach
A merchant or B2B software company might include links to its core commercial pages, key support pages, and trust pages. That could mean the homepage, product overview, pricing, implementation, security, returns or refund policy, company page, and contact page.
What you should not do is stuff the file with every blog post, every tag page, or pages you would not want a prospect to land on. More isn’t better here.
The bigger risk is spending time on llms.txt while missing proven work
Most businesses do not suffer from a missing llms.txt file. They suffer from weaker basics.
Common issues include blocked AI readers, unclear service pages, missing pricing, weak comparison pages, poor review coverage, no author pages, missing local listings, and no citation tracking. Those gaps are far more likely to suppress visibility.
For local firms, the same principle applies. A hotel, restaurant, accountant or solicitor will usually gain more from complete business listings, strong review signals, and direct answer content than from a niche technical file. If local visibility matters to you, see how Galway hotels can appear in AI travel recommendations or how Limerick businesses can appear in AI search results.
The right order of work is access, evidence, content, then optional extras
If you want a sensible decision framework, use this order.
First, check access
Confirm that your AI welcome sign allows the answer crawlers you want to reach your site. Then confirm your server or CDN is not blocking them with a 403 response. This matters because file-level permission and delivery-level permission are not the same thing.
Second, check visibility in first-party tools
Use Google Search Console to review Generative AI impressions by page. Use Bing Webmaster Tools to review citations, cited pages and grounding queries. Use GA4’s AI Assistant channel as a directional traffic source, while remembering it is incomplete by Google’s own published scope.
Third, improve the pages most likely to be cited
Strengthen service pages, product pages, pricing pages, alternatives pages, review pages, comparison pages and support pages. Add direct answers near the top. Add named authors where relevant. Keep dates honest and figures sourced.
Fourth, add optional hygiene items if the basics are already strong
This is where llms.txt belongs. It sits alongside low-cost, checkable items such as your machine-readable business card, called schema markup, and IndexNow for Bing and Copilot workflows. If you need a refresher on that, read what IndexNow is, and which AI assistants it helps.
A simple checklist for deciding whether to add llms.txt
- Check whether your AI welcome sign allows the answer crawlers you want to reach your site.
- Check your CDN or security layer for 403 blocks to AI readers, especially if you use Cloudflare.
- Confirm your key pages are in Google’s phone book and Bing’s phone book.
- Review Google Search Console’s Generative AI report to see which pages already earn AI impressions.
- Review Bing Webmaster Tools AI Performance to see which pages are cited and for which grounding queries.
- Prioritise fixing weak commercial pages, such as pricing, alternatives, product, service and trust pages.
- Add real bylines, honest update dates and sourced figures to pages that need authority.
- If you still want llms.txt, keep it short and link only to your most useful public pages.
- Do not expect llms.txt to replace answer content, reviews, listings or access checks.
- Recheck monthly, because access rules and AI platform behaviour change quickly.
The bottom line is that llms.txt is optional, not urgent
If your house is in order, adding llms.txt is fine. It is a low-cost extra. If your foundations are not in order, it is a distraction.
The businesses that win in AI assistants usually do the boring things well. They stay accessible. They publish pages that answer buying questions directly. They show trust signals clearly. They track what is actually being cited. They don’t chase every new file format as if it were a shortcut.
That is the practical answer for most Irish businesses. Use llms.txt if you want tidy housekeeping. Don’t mistake it for a proven visibility lever.
Is llms.txt an official web standard?
No. It is a proposed convention, not a formal standard adopted across the web. That is one reason to treat it cautiously and keep it in the optional category.
Will llms.txt help my business appear in ChatGPT or Google’s AI answer box?
There is no strong first-party evidence that it will. Your bigger gains usually come from access, clear answer content, trusted commercial pages, reviews, listings and citation tracking.
Can llms.txt replace robots.txt, my AI welcome sign?
No. They do different jobs. Your AI welcome sign controls whether AI readers are invited in. llms.txt is only a suggested signpost to useful pages.
Should local businesses bother with llms.txt?
Usually not as a first step. Local businesses tend to benefit more from complete business profiles, strong review signals, accurate contact details, and pages that answer local buying questions clearly.
If llms.txt is optional, what should I do first?
Start by checking access for AI readers, your presence in Google’s phone book and Bing’s phone book, and your citation data in Google Search Console, Bing Webmaster Tools and GA4. Then improve the pages most likely to be cited.
